Amazon Web Services (AWS) has launched an AI-driven security platform, AWS Continuum, that finds and validates software vulnerabilities and recommends fixes.

SiliconANGLE reported on June 17 that the first feature, a code vulnerability response capability, will be offered first in a limited preview.

Continuum identifies vulnerabilities across customer environments, prioritises them, filters out false positives and suggests remediation. It minimises human involvement but initially runs in a learning mode that includes human review. Customers can later raise the level of automated actions based on defined categories and risk profiles.

AWS says relying only on telemetry data collection, storage and dashboard monitoring by existing corporate security teams makes it hard to keep pace with evolving attacks.

Continuum is not tied to a specific model. It uses multiple AI models that perform better depending on the task. It analyses structured data such as infrastructure, permissions, network topology and code alongside unstructured data such as customer documents, communications and business priorities.

AWS also integrated penetration testing and code scanning features from AWS Security Agent, which it previewed at re:Invent 2025 in December, into Continuum. They will be offered as Continuum Penetration Testing and Continuum Code Scan. It also previewed Continuum Threat Modeling, which generates threat models from design documents or source code and produces results in the STRIDE format.